Institutional Participation, Momentum, and the Thorny Question of When to Exit Winning Trades
My recent efforts to separately define buying and selling power in the stock market were a first effort to capture when markets, in the near term, were more likely to display momentum (continuation of price movement) versus value (reversal of price movement) effects. In a nutshell, I found that momentum to the upside was positively correlated with buying activity (the upticking of a broad range of stocks). Reversals typically followed from high levels of selling activity (the downticking of a broad range of stocks). More recently I've been studying whether buying/selling across a broader range of shares is more predictive of momentum and value than across a smaller number of dominant large caps. More on that topic to come...
What I find interesting about the upticks and downticks, when disaggregated, is that they are typically occurring at precisely the same time. When I look at the upticking and downticking second by second, an unusual number of stocks will tick in the same direction at the same time. This reflects the buying or selling of baskets of stocks, most often either as outright directional bets or to bring futures prices in line with the cash index. Either way, small traders and market makers in individual shares are not typically buying and selling broad baskets of stocks. Such basket execution is a footprint of larger, institutional involvement in the market.
Proceeding on that logic, I constructed a measure of total upticks and downticks on a moment to moment basis. This measure simply looks at the total amount of uptick/downtick movement across stocks and doesn't care whether the ticks are more to the upside or downside. The idea is that more total ticking is a reflection of greater institutional participation. If large (and largely directional) participants are more present in a market, I would expect market moves to have a greater odds of extending. Without such participation, I would expect directional movement to more often run out of gas.
From February, 2012 forward--the period of time in which I assembled moment-to-moment total ticking--I found 165 trading days in which SPY moved more than 50 bps (half a percent) or more to the upside in a trading day (prior day's close to current day's close). Three days later, the average market gain was +.14%, with 107 occasions up and 58 down.
If we simply break down those occasions by median split based on total ticking, the next three days after a high institutional participation winning day averaged a solid gain of +.26% (55 occasions up, 28 down). If the winning day occurred with low institutional participation, the next three days averaged a gain of only +.02% (53 occasions up, 29 down). In other words, days following a solid gain were as likely to rise when institutional involvement was low vs. high, but the degree of follow through was so much greater when institutions were active that essentially all momentum effects (in terms of price movement) occurred at those times.
This is a nice example of the importance of, not only how markets move, but who is in the market. Many valuable research questions follow from this kind of analysis. For instance, does institutional participation early in the day session help predict movement for the remainder of the trading day? Does institutional participation help to predict, not only general market movement, but the movement of individual stocks and sectors? In trading, as in other high performance fields--from cycling to warfare--we increasingly find quantitative tools supporting and informing discretionary decisions. The popular mantra to follow one's trading plans means little if those plans are uninformed.
Further Reading: Factors That Affect Short-Term Stock Market Movement
.
Comments
Post a Comment